ESTIMATES OF GAINS AND LOSSES FOR RESERVOIRS ON THE SNAKE

RIVER FROM BLACKFOOT TO MILNER, IDAHO, FOR SELECTED PERIODS,

1912 TO 1983

By L.C. Kjelstrom

U.S. GEOLOGICAL SURVEY Water-Resources Investigations Report 87-4063

Prepared in cooperation with IDAHO WATER DISTRICT 01

Boise, Idaho 1988 DEPARTMENT OF THE INTERIOR

DONALD PAUL HODEL, Secretary

U.S. GEOLOGICAL SURVEY

Dallas L. Peck, Director

For additional information Copies of this report can write to: be purchased from: District Chief U.S. Geological Survey U.S. Geological Survey, WRD Books and Open-File Reports 230 Collins Road Federal Center, Bldg. 810 Boise, ID 83702 Box 25425 (208) 334-1750 Denver, CO 80225 (303) 236-7476 CONTENTS Page Abstract...... ] Introduction...... 2 Purpose and scope...... 4 Well-numbering system...... 4 Gaging-station numbering system...... 6 Acknowledgments...... 6 Previous work...... 6 Geohydrologic setting...... 8 Gains and losses computed from water budgets...... 9 American Falls Reservoir...... 11 Ungaged inflow...... 13 Reservoir storage...... 29 Reservoir leakage...... 31 Lake Walcott...... 38 Milner Lake...... 44 Gain and loss analysis...... 44 Reservoir storage...... 54 Summary...... 53 References cited...... 61

ILLUSTRATIONS

Figure 1. Map showing study area...... 3 2. Diagram showing well-numbering system...... 5 3. Map showing data-collection sites near American Falls Reservoir...... IQ 4-13. Graphs showing: 4. Daily inflow from direct precipitation on American Falls Reservoir...... 14 5. Daily rates of evaporation from American Falls Reservoir...... 15 6. Annual mean ground-water discharge to American Falls Reservoir, 1912-83..... 17 7. Ungaged surface-water inflow sources to American Falls Reservoir...... is 8. Differences between average monthly ungaged inflow to the Snake River between Blackfoot and Neeley before (1912-25) and after (1951-82) construction of American Falls Dam.... 20 9. Frequency curve for regression residuals of ungaged inflow to American Falls Reservoir as a per­ centage of combined monthly dis­ charge of the Snake River at Black- foot and at Neeley, 1952-82...... 25 ILLUSTRATIONS Continued Page Figure 10. Differences between three regression equation estimates of monthly mean ungaged inflow to American Falls Reservoir and ungaged inflow as compiled by water budgets...... 28 11. Increases in daily discharge of Spring Creek above base flow and recorded precipitation at Fort Hall and Pocatello, March 1 to September 30, 1983...... 30 12. Comparison between water-budget gain or loss and estimated ungaged inflow to American Falls Reservoir based on Spring Creek discharge, March 1 to September 30, 1983...... 33 13. Standard errors of two daily water- budget estimates of ungaged inflow to American Falls Reservoir from estimates derived from regression equations that relate ungaged inflow to Spring Creek discharge...... 34 14. Map showing Lake Walcott reach of the Snake River...... 35 15-19. Graphs showing: 15. Stage of American Falls Reservoir and elevation of water levels in down- gradient wells...... 36 16. Stage of Bonanza Lake and water levels in nearby wells...... 37 17. Comparison between 5-year moving average gain to Lake Walcott and cumulative departures from mean discharges of Goose Creek near Oakley, 1953-83...... 39 18. Statistical summary of gains to Lake Walcott during low flow in the Snake River, 1935-82...... 43 19. Map showing Milner Lake reach of the Snake River...... 45 20-27. Graphs showing: 20. Outflow from Milner Lake...... 46 21. Mean monthly gains and losses in Milner Lake reach, 1975-83...... 47 22. Standard deviations of gains or losses in Milner Lake and possible discharge- measurement errors in mean monthly discharge, 1975-83...... 43 23. Water levels in well 9S-25E-23DBA1, water year 1983...... 51

11 ILLUSTRATIONS Continued Page Figure 24. Five-year moving average of annual mean gain in Milner Lake and water levels in well 9S-25E-23DBA1...... 52 25. Monthly mean gain or loss in Milner Lake for selected months, 1977-82..... 53 26. Water-surface profiles in Milner Lake for a constant reservoir elevation at Milner and two discharges...... 55 27. Water-surface profiles in Milner Lake for a constant inflow and two reser­ voir elevations at Milner...... 56

TABLES

Table 1. Ungaged surface-water inflow to American Falls Reservoir, 1980...... 12 2. Average monthly ungaged inflow to American Falls Reservoir and corresponding standard deviations, 1951-82...... 19 3. Monthly and annual water budgets for water year 1983 used to determine ungaged inflow to American Falls Reservoir...... 21 4. Sources and amounts of ground-water recharge supplying water for spring discharge...... 23 5. Comparison of standard errors of gains com­ puted from combinations of reservoir stage gages (April 15 to September 15, 1983)...... 32 6. Inflow to Lake Walcott, 1951-83...... 41 7. Surface-water inflow to the Snake River between Minidoka and Milner...... 50 8. Storage-capacity table for Milner Lake...... 57

111 CONVERSION FACTORS For the convenience of readers who may prefer to use metric (International System) units rather than the inch- pound units used in this report, conversion factors are listed below: Multiply To obtain inch-pound unit By metric unit

acre 4,047 square meter acre-foot (acre-ft) 1,233 cubic meter cubic foot per second 0.02832 cubic meter (ft3 /s) per second foot (ft) 0.3048 meter inch (in.) 25.40 millimeter megawatt-hour 3,600,000,000 joule (MWh) mile (mi) 1.609 kilometer square mile (mi2 ) 2.590 square kilometer Sea level ; In this report, "sea level" refers to the National TSeodetic Vertical Datum of 1929 (NGVD of 1929) a geodetic datum derived from a general adjustment of the first-order level nets of both the United States and Canada, formerly called "Mean Sea Level of 1929."

IV ESTIMATES OF GAINS AND LOSSES FOR RESERVOIRS

ON THE SNAKE RIVER FROM BLACKFOOT TO MILNER, IDAHO,

FOR SELECTED PERIODS, 1912 TO 1983 By L.C. Kjelstrom

ABSTRACT

Croplands in the semiarid central part of the Snake River Plain are dependent on the availability of irrigation water, most of which comes from the Snake River. Allocation of irrigation water from the river requires that gains and losses be determined for American Falls Reservoir, Lake Walcott, and Milner Lake. From 1912 to 1983, average ungaged inflow to American Falls Reservoir, determined from monthly water budgets, was 2,690 cubic feet per second. About 94 percent of this inflow was spring discharge and ground-water seepage; the remainder was from small tributaries and irrigation-return flow. Ungaged inflow estimated from water budgets for various periods correlated favorably with measured discharge of two springs and water levels in two wells. Discharge of Spring Creek was a better indicator of ungaged inflow than ground-water levels. Therefore, correlation with Spring Creek discharge was used in estimating ungaged inflow to American Falls Reservoir in 1983. Daily water-budget calculations of ungaged inflow to American Falls Reservoir are less variable when storage changes are determined by using three stage-recording stations rather than one. Water budgets do not indicate large amounts of leakage from American Falls Reservoir, but small amounts of leakage are indicated because flow in downstream springs increased about 25 percent after reservoir storage began in 1926. Water budgets for Lake Walcott and Milner Lake show average annual net gains (1951-83) to Lake Walcott and Milner Lake of 245 and 290 cubic feet per second. These amounts are verified by (1) monthly water budgets when discharge in the Snake River is low, and (2) measured and estimated sources of inflow. Gains and losses estimated from daily water budgets are variable, owing to inadequate determination of (1) changes in reservoir storage, (2) streamflow, (3) lake-surface precipitation, and (4) lake- surface evaporation. Backwater effects are accounted for in the process used to determine storage in Milner Lake. INTRODUCTION The economy of the semiarid Snake River Plain, southern Idaho, relies heavily on irrigated agriculture. American Falls Reservoir, Lake Walcott, and Milner Lake (fig. 1) supply irrigation water to the central area of the plain. Diversions are made to fulfill established water rights on natural streamflow (without regulation or diversion) and reservoir storage. To determine the amount to be diverted, which depends on the amount available, water managers use water budgets to compute gains to and losses from the Snake River between U.S. Geological Survey gaging stations near Blackfoot and at Neeley (above and below American Falls Reservoir), at Neeley and near Minidoka (above and below Lake Walcott), and near Minidoka and near Milner (above and below Milner Lake). During the 1983 irrigation season (April-October), diversions from American Falls Reservoir, Lake Walcott, and Milner Lake were about 42,900, 694,500, and 2,597,300 acre-ft. This water is used to irrigate about 600,000 acres (Water District 01, 1981). The daily gain or loss in each reservoir is presently (1983) computed by the Snake River watermaster by use of a water budget, where: gain or loss = measured outflow (to river and canals) - measured inflow +^ storage changes + evaporation from the reservoir Differences between discharges of the river at gaging stations upstream and downstream from each reservoir (ad­ justed for change in storage, diversions, and evaporation) are considered to constitute the net gain or loss for each reservoir. Gains include ungaged discharge from springs and small tributaries, ground-water seepage, irrigation-return flow, and precipitation on the reservoirs. Losses include percolation to ground water and bank storage. Gain and loss values determined by water budgets include an error component composed of discharge-measurement errors, change-in-storage errors, and undefined interactions between ground and surface water. Inaccuracies may be reduced by collection of additional data. Where additional data collection is not possible or where the quality of the data is not likely to reduce inaccuracies, gains and losses in the reservoirs may be more accurately defined by re- examining available data and the relations that exist between ground water and surface water. 'AMERICAN FALLS RESERVOIR Lake Channel American CO Bonanza Falls Dam Neeley

EXPLANATION A GAGING STATION

30 MILES Base from U.S. Geological Survey State base map 1:1,000,000 0 10 20 30 KILOMETERS

Figure 1. Study area. Purpose and Scope The purpose of this report was to present results of a study to: (1) Determine, by use of water budgets, gains and losses for three Snake River reservoirs between Blackfoot and Milner; (2) account for budget errors; and (3) suggest ways to reduce or eliminate those errors. The major study effort was focused on the river reach between the gages near Blackfoot and at Neeley, which includes American Falls Reservoir. This was done because of the size of the reservoir and the effect that unmeasured ground-water inflow appears to have on the water budget. Analyses of gains and losses determined from water budgets were made by using available data and data collected for this study. This study was made in cooperation with Idaho Water District 01 during water years 1983 and 1984. The objective of the study was to provide information for the management of irrigation diversions, reservoir storage, and flow in the Snake River. The scope of the study included (1) collecting water- stage data at three sites in American Falls Reservoir, two sites in Milner Lake, and one site in Bonanza Lake (a landlocked lake) from November 1982 to September 1983; (2) computing daily change in storage by using the three stage gages in American Falls Reservoir; (3) operating a discharge-measurement station on Spring Creek at Sheepskin Road near Fort Hall from August 1980 through September 1983; (4) measuring water levels in 10 wells from November 1982 through September 1983 (3 wells were equipped with contin­ uous recorders and 7 were measured monthly); and (5) devel­ oping a storage-capacity table for Milner Lake that con­ siders backwater conditions. Existing data used in the study included discharge records for four sites on the Snake River and one site on the Portneuf River, stage records from gages near the dams of all three reservoirs, diversion data, and records from five recording precipitation stations and from one pan- evaporation station.

Well-Numbering System The well-numbering system (fig. 2) used by the U.S. Geological Survey in Idaho indicates the location of wells within the official rectangular subdivision of the public lands, with reference to the Boise base line and Meridian. For example, well 5S-31E-27ABA1 is in the NE^NW^NE^, sec. 27, T. 5 S., R. 31 E. The first two segments of the number designate the township (5S) and range (31E). The third 5S-31E-27ABA1

Figure 2. Well-numbering system. segment gives the section number (27); three letters (ABA), which indicate the k section (A, 160-acre tract), h~h section (B, 40-acre tract), %-%-% section (A, 10-acre tract); and sequence number of the well (1) within the tract, if available. Quarter sections are lettered A, B, C, and D in counterclockwise order from the northeast quarter of each section. Within the quarter sections, 40-acre and 10-acre tracts are lettered in the same manner. The se­ quence number indicates it was the first well inventoried in that tract.

Gaging-Station Numbering System Each continuous- and partial-record gaging station in Idaho has been assigned a number in accordance with the permanent numbering system used by the U.S. Geological Survey. Numbers are assigned in downstream order along the main stream, and stations on tributaries between main-stream stations are numbered in the order they enter the main stream. A similar order is followed on other ranks of tributaries. The complete eight-digit number, such as 13075500, which is used for the station "Portneuf River at Pocatello," includes the part "13," indicating that the Portneuf River is in the Snake River basin, plus a six-digit number.

Acknowledgments The cooperation and technical contributions of the following individuals and groups are greatly appreciated. Mr. A. C. Robertson and Mr. R. Sutter of the Idaho Depart­ ment of Water Resources provided technical advice and tabulated data. The Shoshone-Bannock Tribal Council and other landowners allowed access to gaging stations and wells, which greatly facilitated the study. The U.S. Bureau of Reclamation cooperated in obtaining recorded-stage data on the reservoirs. Mr. Jerry Eggleston of CH2M Hill Engi­ neering furnished cross-sectional data for Milner Lake.

PREVIOUS WORK Reports by Newell (1928, 1929) and Mundorff (1967) provide detailed discussion and quantitative analysis of ungaged ground- and surface-water inflow to American Falls Reservoir. Inflow to Lake Walcott is discussed in detail in reports by Meisler (1958) and Mundorff and others (1964). After American Falls Dam was completed and reservoir storage began in 1926, irrigation water was allotted by the watermaster. To differentiate natural flow to the reach from reservoir storage, Newell (1928, 1929) measured flow of 27 tributaries largely discharge from springs to American Falls Reservoir during the irrigation seasons of 1927 and 1928. Combined discharge of the 27 sites was used as an index of total ungaged inflow (obtained by water- budget analysis) and the following relation was determined: I = 840 + 1.33 Qm (1) where I = ungaged inflow, in cubic feet per second, 840 = discharge from less variable sources, and Qm = total discharge at 27 measuring sites, in cubic feet per second. Ungaged inflow to the reservoir was estimated by the water- master using this relation during each succeeding irrigation season from 1932 to 1977. Since 1977, ungaged inflow has been computed as the residual of daily water budgets or the remainder after subtraction of measured surface inflow (largely from springs) from measured outflow with adjust­ ments for storage change and evaporation. Newell (1928, 1929) measured water levels in wells, estimated reservoir evaporation from pan-evaporation data, and determined bank storage and reservoir-seepage losses for American Falls Reservoir. Profiles he constructed from water-level data showed that the water table in the shallow aquifer generally sloped toward the reservoir except at the southwestern end of the reservoir, where ground-water levels were generally below those of the reservoir water surface. Mundorff (1967) studied the effects on ground-water levels and spring discharges that would result from raising the height of American Falls Dam 15 ft. He estimated that ground-water discharge from springs in the early 1900's was about 1.2 to 1.4 million acre-ft/yr, but that seepage losses from the reservoir, built in 1926, were negligible because fine-grained deposits that crop out along the shore and extend to depths of 50 to 100 ft form an effective seal beneath the reservoir. Meisler (1958) investigated the area north and west of American Falls Dam to determine the direction of ground- water movement and the relation of the regional ground-water system to the Snake River. He concluded that ground-water movement in the Bonanza Lake area is generally southwestward to the Snake River or westward toward the Snake Plain aquifer and that ground water in the Bonanza Lake area may be perched above relatively impermeable strata, perhaps lakebeds of the Raft Formation (Meisler, 1958, p. 18, 19). Stearns and others (1938, p. 137-142) concluded that recharge to springs discharging to American Falls Reservoir is mostly underflow from the Snake River far upstream. Ground-water underflow from the Portneuf River valley and precipitation on lava beds north of the Snake River were suggested as other sources of recharge. After American Falls Reservoir was filled in 1926, Stearns and others (1938, p. 154) observed a rise in the level of Bonanza Lake, southwest of American Falls Dam. They concluded that this rise, along with the hydraulic gradient determined from water-level measurements in wells, indicated leakage from the reservoir moves southwestward and affects stage of Bonanza Lake and discharge of several springs on the north side of Lake Walcott. Crosthwaite (1974, p. 17) suggested that hydraulic gradients in alluvium and basalt overlying the Raft Forma­ tion in the general area of Lake Channel slope southward. His data from a resistivity profile and an exploratory well (8S-29E-34CBC1) indicated that basalt of the Snake River Group is generally above the regional water table from the vicinity of Bonanza Lake southward to the Snake River. Water moving southwestward through basalt east and northeast of Lake Channel is perched above the Raft Formation.

GEOHYDROLOGIC SETTING Basaltic rocks with lesser amounts of intercalated, unconsolidated sedimentary rocks compose the Snake River Plain aquifer system. Permeable zones between adjacent lava flows result in high aquifer transmissivities. Fine-grained sediments, exposed south of and underlying American Falls Reservoir, grade into sand and gravel layers interbedded with basalt west of the reservoir. Large springs have developed at the upper end of the reservoir where tributaries have cut downward through fine-grained sediments into gravels. Lake Channel, southwest of American Falls Dam, is an abandoned channel of the Snake River that is partially filled with basalt flows and fine sediments. Bonanza Lake and several other spring-fed ponds are located in Lake Channel. From Bonanza Lake to the Snake River, loess and alluvium overlie basalt flows and fine sediments. Lakebeds of the Pleistocene Raft Formation along the south side of American Falls Reservoir and upper Lake Walcott probably extend north of the river under basalt flows. Basalt flows with interbedded sediments extend westward and downstream on both sides of the Snake River from Lake Channel on the north and Raft River on the south. In some areas north of Milner Lake, alluvial deposits overlie the basalts and interbedded sediments. Water-table gradients west and north of American Falls Reservoir are toward the reservoir from a ground-water divide (Mundorff, 1967, p. 15) approximately parallel to the Aberdeen-Springfield High Line Canal (fig. 3). Because the relatively impervious Raft Formation along the western shore of the reservoir restricts ground-water movement, the water table is generally above reservoir high-water stages. Comparison of historical and recent water-table contour maps (Garabedian, 1986, pi. 1) shows the configura­ tion of the water table has remained relatively unchanged since 1928. Seasonal fluctuations of the water table are caused by irrigation practices and differ depending on the local effects of recharge from surface-water irrigation or discharge from pumpage (Young and Norvitch, 1984, p. 5-7). Mundorff (1967, p. 30) showed a correlation between water levels in well 5S-31E-27ABA1 (fig. 3) and monthly ground-water discharge to American Falls Reservoir estimated from water-budget analysis. When water levels are high, estimated ground-water discharge to American Falls Reservoir tends to be high, and when water levels are low, estimated ground-water discharge tends to be low.

GAINS AND LOSSES COMPUTED FROM WATER BUDGETS Gains and losses in American Falls Reservoir, Lake Walcott, and Milner Lake were computed from annual, monthly, and daily water budgets. Specific methods and data used to compute the water budgets are discussed in following sec­ tions. Annual water budgets were used to indicate overall trends in gains and losses and to determine annual averages; monthly water budgets were used to compare gains and losses with a selected range of discharge-measurement accuracy and to determine monthly averages; daily water budgets were used to show errors in determining short-term changes in reser­ voir storage, precipitation, and evaporation. Sources of gains to and losses from each reservoir were quantified on the basis of available data. Gains determined by water-budget analysis were compared with estimations of ungaged inflow. Losses estimated by water budgets were compared with water-level changes and water levels in relation to reservoir stage. Benefits of ad­ ditional data collection to determine gains and losses are discussed in the following sections. 112°30'

113°00

EXPLANATION

STAGE-MEASUREMENT STATION

DISCONTINUED GAGING STATION

GAGING STATION

PRECIPITATION RECORDING STATION

OBSERVATION WELL AND NUMBER

42°45' 10 MILES 10 KILOMETERS

Figure 3. Data-collection sites near American Falls Reservoir. American Falls Reservoir American Falls Reservoir, when at a maximum operating elevation of 4,354.5 ft above sea level, has a surface area of nearly 60,000 acres. Usable reservoir storage is about 1.7 million acre-ft between normal operating elevations 4,295.7 and 4,354.5 ft. American Falls Dam was completed in 1926 and reconstructed in 1977. Stored water from American Falls Reservoir is restored in Lake Walcott and Milner Lake where it is diverted for irrigation. Storage space in American Falls Reservoir can also be used for flood control. A powerplant at the dam generates about 340,000 MWh annually (Heitz and others, 1980, p. 130). Gain to American Falls Reservoir was determined for this study as: Gain = QN + D + E + SC - QB - Qp - R (2) where Q = discharge of the Snake River at Neeley, D = discharge diverted for irrigation, E = evaporation from American Falls Reservoir, SC = change in reservoir storage, Q_ = discharge of the Snake River near Blackfoot, D Q = discharge of the Portneuf River at Pocatello, and R = precipitation on reservoir water-surface area. This computed gain represents the sum of the unmeasured ground- and surface-water inflow and hereafter is referred to as ungaged inflow. Ungaged inflow (gain) includes discharge from springs and small tributaries, ground-water seepage, and irrigation-return flow. Inflow from these sources accounts for nearly 35 percent of total inflow to American Falls Reservoir (table 1). The remaining 65 percent is measured discharge in the Snake River near Blackfoot and Portneuf River near Pocatello (fig. 3). Inflow to the reservoir from precipitation was esti­ mated by using daily mean rainfall data from weather sta­ tions at Aberdeen, American Falls, Pocatello, and Blackfoot (National Oceanic and Atmospheric Administration, 1983). On the basis of a gage density equivalent to 250 mi 2 per gage, Winter (1981, p. 87) showed that by using these data, errors up to 60 percent can be introduced. A rectangular area that includes American Falls Reservoir and the above four weather stations covers about 400 mi 2 ; within that rectangle, gage density is about one per 100 mi 2 . However,

11 Table 1. Ungaged surface-water inflow to American Falls Reservoir, 1980 [Site locations shown on figure 7]

Average inflow Site (cubic feet No. Source per second)

1 Crystal Wasteway l 37.0 2 Sterling Wasteway 1 14.6 3 Coburn Wasteway 1 5.4 4 Aberdeen Wasteway 1 28.8 5 Pocatello Creek 2 4.7 6 Fort Hall Main Canal waste 3 3.2 7 Ross Fork l 60.6 8 Dubois lateral waste 3 1.1 9 Tyhee lateral waste 3 2.7 10 Church lateral waste 3 3.8 11 Gibson drain 3 2.8 12 Bannock Creek 1 48.2 13 Tarter drain 1 6.1 14 Schiltz drain 1 3.6 15 Tributary to Seagull Bay 1 .3 16 Sunbeam Creek 1 1.5 17 Cedar Creek l .4 Total 224.8

1 Average of bimonthly measurements made in 1980. 2 Estimate from regional regression equation (Kjelstrom, 1986). 3 Data from U.S. Bureau of Reclamation gages in 1980.

12 increasing gage density would not result in appreciably improved accuracies because mathematical methods used to calculate areal averages differ by as much as 18 percent (Winter, 1981, p. 86), and instrument errors and errors caused by placement of gages would remain. Precipitation on the reservoir equals about 0.5 percent of the total average annual inflow. Most rain falls during only a few days each year; however, on those particular days, rainfall can have a significant impact on water budgets. For instance, 1 in. of rain falling on the water surface of the reservoir in 24 hours is equivalent to an inflow of about 3,200 acre-ft, or 1,600 ft 3 /s. Daily inflow to the reservoir from precipitation for 5 months in 1983 is shown in fig. 4. On 13 days between May and September 1983, daily inflow from precipitation on American Falls Reservoir was greater than 400 ft 3 /s. Outflow from American Falls Reservoir consists of discharges of the Snake River at Neeley, two irrigation diversions, and evaporation from the reservoir. In 1983, discharge of the Snake River at Neeley ranged from 91 percent of daily outflow from the reservoir in the summer to nearly 100 percent in the winter. The average flow at Neeley is 7,260 ft3 /s and in 1983, flows ranged from 832 to 30,700 ft3 /s. Average flow for water year 1983 was 11,740 ft 3 /s. An average of 175 ft s /s was diverted for irrigation for about 120 days between May and September 1983. Reservoir evaporation was estimated using pan data collected at the University of Idaho's Aberdeen Experiment Station (National Oceanic and Atmospheric Administration, 1983). A coefficient of 0.70 was used to convert pan evaporation to reservoir evaporation. Rates of evaporation are variable (fig. 5), and daily estimates of evaporation may introduce significant errors in daily water budgets. These errors probably would remain relatively high, even with additional data collection and use of more sophisti­ cated estimation techniques. Average annual evaporation from American Falls Reser­ voir from 1946 to 1958 was about 38 in. (Meyers, 1962, p. 94), or 180,000 acre-ft. Total evaporation from April to October 1983 was nearly 34 in. (160,000 acre-ft), or about 90 percent of average annual evaporation for the period 1946-58.

Ungaged Inflow Annual mean ground-water discharge, which accounts for about 94 percent of the average annual ungaged inflow to American Falls Reservoir, has been relatively steady from

13 1,UUU 1 1 1

- - Q z S 1.200 - - UJ

Of UJ - - 1-0. LU UJ " 800 - - y CQ D U Z

O 400 - - u. Z

0 | | i i 1 1 May ' June 1 July August ' September ' 1983

Figure 4. Daily inflow from direct precipitation on American Falls Reservoir. c"-* EVAPORATION, IN INCHES PER DAY (@

3

< o 2.2 n A o s Oct. "-* o B 1912 to 1983 (fig. 6). Water-budget analysis indicates that average annual ungaged inflow to American Falls Reservoir from 1912 to 1983 was about 2,690 ft 3 /s. On the basis of miscellaneous discharge measurements and water yield esti­ mates from 1934 to 1980 (Kjelstrom, 1986), surface-water inflow from small tributary streams and irrigation-return flow account for an average annual discharge of about 150 ft 3 /s. The ratio of annual to average annual discharge of the Portneuf River was used to vary the average annual discharge from small tributaries for each year. The per­ centage of diverted irrigation water that returned to the reservoir each year was assumed to be equivalent to the percentage obtained from measurements made in 1980. The sources and relative volumes of ungaged surface- water inflow to American Falls Reservoir (fig. 7) are given in table 1. The higher-than-average (1912-83) total dis­ charge may be attributed to above-normal precipitation, base or ground-water flow being included at several of the sources, and estimation errors. In addition, Spring and Clear Creeks, although primarily fed by springs, collect some runoff from adjacent areas. Mean monthly ungaged inflow was determined from water budgets for the period 1951-82; averages and corresponding standard deviations for each month are given in table 2. An examination of differences between monthly ungaged inflow for the period 1951-82, when most surface-water irrigated land had been developed, and for the period 1912-25 (also determined by water-budget analysis), prior to construction of American Falls Dam (fig. 8), shows that the post-dam and post-irrigation development inflow is greater from June to January and less from March to May than pre-dam and pre- irrigation development inflow. No climatological change is indicated when nearly similar periods (1912-25 and 1962-80) of mean monthly flows of the Snake River near Heise adjusted for upstream storage are compared (Kjelstrom, 1986). The increase may be attributed to seasonal rises in ground-water levels owing to irrigation and additional water released from bank storage. The decrease may be attributed to loss of water to bank storage when the reservoir stage is rising. For water year 1983, monthly mean ungaged inflows range from 2,330 to 3,260 ft 3 /s (table 3). The annual mean ungaged inflow for water year 1983 is about 2,800 ft 3 /s. Annual mean surface-water inflow, based on discharge from the Portneuf River and upstream diversions, was esti­ mated as 176 ft 3 /s. Ground-water discharge was assumed to be the residual, or about 2,620 ft 3 /s. Ground-water discharge, primarily from springs, is associated closely with the regional ground-water system and, as would be expected, ground-water levels correlate

16 3,000 IT IT TT IT TT TT TT u 50 "I

2,500 K QL

U UJ

Q

2,000 I I I I I I 11 1 1910 1920 1930 1940 1950 1960 1970 1980 1985 WATER YEAR

Figure 6. Annual mean ground-water discharge to American Falls Reservoir, 1912-83. 112°30

113°00

43°00' EXPLANATION

[3J SURFACE-WATER INFLOW MEASURING SITE AND NUMBER-Site numbers correlate with table 2 oo

42°45' 10 MILES 10 KILOMETERS

Figure 7.--Ungaged surface-water inflow sources to American Falls Reservoir. Table 2. Average monthly ungaged inflow to American Falls Reservoir and corresponding standard deviations, 1951-82

Standard Average monthly deviation Standard ungaged inflow (cubic deviation (cubic feet feet (percent Month per second) per second) of mean)

October 2,920 180 6.2 November 2,750 180 6.5 December 2,610 210 8.0 January 2,600 180 6.9 February 2,570 270 10.5 March 2,520 260 10.3 April 2,430 380 15.6 May 2,490 310 12.4 June 2,790 340 12.2 July 2,930 320 10.9 August 3,050 180 5.9 September 2,890 220 7.6

19 *1 ^w C/5 C 3 *

i^t QO

VD v. L/l D O D O O 1 1 1 00 S ^ Oct-

^^ .. O Nov. - ^ BM -^ °P JrftT

0 pa 2 Mar. 3 2 <» 0 ^ » Apr.

> g Aug.

^ 3 pa Sept. 1 1 1 1 1 1

5fl ^* rs . ^O ^^

Q 1 * N P9 Ki 3 31 ^^ ^ 5"

3 0 3-

Monthly mean discharge (cubic feet per second)

Diversions Evaporation Change Snake Portneuf Snake River for from in River near River at Precipitation Ungaged at Neeley irrigation reservoir storage Blackfoot Pocatello on reservoir inflow Month ( + ) (+) (+) ( + or -) (-) (-) (-) (residual) D E SC R % QB OP

October 6,938 0 150 +2,090 5,919 340 70 2,850 November 9,644 0 80 -300 6,026 336 100 2,960 to December 10,320 0 50 -1,120 5,943 314 110 2,870 January 9,600 0 40 +390 6,700 377 30 2,920 February 7,249 0 60 +1,440 5,272 410 70 3,000 March 4,655 0 110 +4,820 5,936 756 150 2,740 April 11,210 0 290 +1,070 9,058 875 30 2,610 May 22,330 16 410 -30 18,440 1,513 150 2,620 June 22,810 206 450 +190 19,150 1,142 100 3,260 July 16,210 245 490 -1,340 12,830 363 80 2,330 August 10,420 110 410 -3,200 4,641 290 100 2,710 September 9,253 112 380 -3,820 2,760 338 100 2,730

Annual 11,740 58 240 +12 8,576 588 90 2,800 with spring discharge to American Falls Reservoir. Table 4 shows sources and amounts of ground-water recharge that may supply springs. For 1983, recharge was balanced with discharge by seepage of irrigation water on land upstream from American Falls Reservoir. Prior to irrigation on the Snake River Plain, recharge that supplied springs in the Blackfoot to Neeley reach was primarily from upstream channel losses in the Snake River. Tributary underflow to and precipitation on the land being recharged by irrigation water are assumed to have supplied about the same amounts both before and after irrigation and are based on 1934-80 averages of water yield and precipitation (Kjelstrom, 1986). Water-budget analysis shows that average annual losses from the Snake River between Heise and Blackfoot from 1912 to 1922 were 750,000 acre-ft. A water-budget analysis for 1974-80 shows losses of 280,000 acre-ft. The decrease in losses can be attributed to less flow in the Snake River as a result of increased upstream diversions and, possibly, high water tables where the river is hydraulically connected with ground water. From the early 1890's to the late 1950's, when most of the surface-water irrigated land was developed, the regional water table rose about 60 to 70 ft (Mundorff and others, 1964, p. 162), and ground-water discharge as spring flow to the Snake River from Blackfoot to Neeley nearly doubled (table 2). From this general correspondence between rises in the water table and increases in spring flow, it is reasonable to expect that water levels in particular wells could be related to spring flow. Well 5S-31E-27ABA1 (fig. 3) is drilled in basalt of the Snake River Group to a depth of 50 ft. Mundorff (1967) showed that hydrographs of water levels in this well correlated with discharge hydrographs for nearby springs. To determine the current degree of correla­ tion, mid-month water levels in well 5S-31E-27ABA1 from May 1952 to September 1982 were regressed with monthly mean ungaged inflows. After the months with missing data were eliminated, 346 months were available for regression analy­ sis. The RMSE (root mean square error) determined by using the regression was 290 ft 3 /s, which is about 10 percent of the average annual ungaged inflow of 2,740 ft 3 /s. The RMSE is the standard deviation of the distribution (assumed normal) of residuals about the regression line. Residuals are the difference between ungaged inflows determined by water budget and those determined by the regression. Storage changes in American Falls Reservoir also may be a source of error in the water budget. While the dam was being reconstructed in 1977 and the reservoir was empty, the reservoir basin was resurveyed and a new capacity table was developed (U.S. Bureau of Reclamation, 1979). To determine errors owing to storage change on regression residuals, water budgets that determined ungaged inflow were recomputed

22 Table 4. Sources and amounts of ground-water recharge supplying water for spring discharge

Relative amounts of recharge (millions of acre-feet) Sources of ______ground-water recharge 1983 Prior to irrigation

Irrigation losses 1.4 0 Snake River losses .3 .8 Tributary underflow .1 .1 Precipitation .1 .1 Total or ground-water discharge 1.9 1.0

23 using the new capacity table for the 20 months with the largest residuals. The resulting regression equation reduced the RMSE from 290 to 260 ft 3 /s. From this analysis, it appears that the original reservoir capacity table may account for a small part of the error. Although errors in determining change in reservoir contents for monthly water budgets appear to be small, possible errors in daily water budgets can be large. An error of only 0.01 ft of reservoir stage results in a difference in estimated ungaged inflow of 250 to 500 ft 3 /s, depending on the water-surface area of the reservoir. However, errors in defining reservoir stage that are due to wind or wave effects (and lag times) are evened out by using longer periods of record or monthly water budgets. An evaluation of the discharge records may be made by analysis of differences between estimates of ungaged inflow determined by a water budget and those determined by regres­ sions based on ground-water levels. About 94 percent of the differences are less than 5 percent of the combined Snake River discharges near Blackfoot and at Neeley (fig. 9). Combined discharge more closely represents actual flow through the reservoir system than inflow or outflow. Although inflow is high in the spring, outflow may be small while the reservoir is being filled. In late summer, inflow may be small because of upstream diversions and outflow may be large because of releases for downstream diversions. Discharge records are considered excellent if about 95 percent of the daily mean discharges are within 5 percent of the true discharge. Although monthly mean discharges would be more accurate than daily mean discharges because of the smoothing effect of the longer period, the distribution of differences appears to be nearly within the error range of excellent discharge records. The relatively low discharge- measurement errors may be attributed to the basalt channel of the Snake River that provides stable controls for the stage-discharge relation. Regression estimates of ungaged inflow may avoid the large variations in water-budget estimates that could be attributed to discharge-measurement errors and that are not representative of the relatively steady inflow from 1912 to 1983, mostly from ground water. In addition, other inde­ pendent variables may be used in the regression. Miscel­ laneous measurements of Spring Creek discharge from 1926 to 1978 (U.S. Geological Survey, 1968-78; Decker and others, 1970), averaging about 20 percent of the ungaged inflow to the reservoir (most of which is spring discharge), correlate well with ungaged inflow computed by Newell's formula (C.A. Thomas, U.S. Geological Survey, oral commun., 1980). A

24 10

9

u. o z

^^ £5 Sf slK Z

10 20 30 40 50 60 70 80 90 100 PERCENTAGE OF RESIDUALS LESS THAN INDICATED VALUE

Figure 9. Frequency curve for regression residuals of ungaged inflow to American Falls Reservoir as a percentage of combined monthly dis­ charge of the Snake River at Blackfoot and at Neeley, 1952-82. gaging station was established in August 1980 on Spring Creek at Sheepskin Road near Fort Hall, about 5 mi upstream from the miscellaneous measurement site used from 1926 to 1978. The station was established there to avoid backwater conditions from the reservoir. Flow past the gaging station averages about 15 percent of the ungaged inflow to American Falls Reservoir. On the basis of several sets of discharge measurements, ground-water inflow between the two sites ranges from 90 to 120 ft3 /s. Monthly mean discharges of Spring Creek from August 1980 to September 1982 were regressed with monthly ungaged inflow. Mid-month water levels in well 5S-31E-27ABA1 also were regressed with ungaged inflow for the same period. The RMSE's were 332 ft 3 /s for ungaged inflow estimated from discharge in Spring Creek and 294 ft3 /s estimated from water levels in the well. Regression of ungaged inflow with discharge in Daniel- son Creek (fig. 3) for 14 months from August 1980 to September 1981 resulted in an RMSE of 322 ft 3 /s. When both Danielson Creek (spring fed) and Spring Creek discharges were used as independent variables, the RMSE was reduced to 264 ft 3 /s. The gaging station on Danielson Creek was discontinued in 1981 so was not used in further regression analysis. Discharge records for the Snake River near Blackfoot and at Neeley were reanalyzed for 12 of the 26 months from August 1980 to September 1982 when residuals from Spring Creek discharge or ground-water level regressions exceeded the RMSE. Changes in shift adjustments applied to the stage/discharge relation during 5 months at Neeley and 1 month near Blackfoot resulted in increases or decreases of 1 to 3 percent in monthly mean discharges. Ungaged inflow was recomputed using the revised discharge records at Blackfoot and Neeley. These ungaged inflows were related to Spring Creek discharge by: Q = 2,140 + 6.90 (Qs) (3) where Q = ungaged inflow to American Falls Reservoir, in cubic feet per second; and Qs = Spring Creek discharge, in cubic feet per second minus 250 ft3 /s (for easier computation of the regression equation). The equation that resulted from regressing with the revised ungaged inflow on mid-month water levels of well 5S-31E- 27ABA1 is:

Q = 2,130 + 55.8 (L) (4)

26 where Q = ungaged inflow to American Falls Reservoir, in cubic feet per second; and L = water-surface elevation, in feet above sea level, minus 4,370 ft (for easier computation of the regression equation). The RMSE was 196 ft 3 /s. Multiple regression of both Spring Creek discharge and water levels in well 5S-31E- 27ABA1 as independent variables did not decrease the RMSE. Another possible variable for regression with ungaged inflow is water level in well 4S-33E-3CBB2, located about 7 mi north of the reservoir and completed in basalt of the Snake River Group. Water levels in the well were recorded continuously from 1959 to 1969 and were measured periodi­ cally during 1970-77. A recorder was reinstalled in January 1978. Water levels from 1978 to 1982 were lower than during 1959-69 (probably as a result of a drought in 1977), with fewer seasonal fluctuations than during the 1959-69 period. The equation that resulted from regressing mid-month water levels of well 4S-33E-3CBB2 with ungaged inflow is:

Q = 1,230 -I- 164 (L) (5) where Q = ungaged inflow to American Falls Reservoir, in cubic feet per second; and L = water-surface elevation, in feet above sea level, minus 4,440 ft (for easier computation of the regression equation). The RMSE of a regression with ungaged inflow into American Falls Reservoir for the period August 1980 to September 1982 was 185 ft3 /s. Equations (3), (4), and (5) were used to estimate monthly mean ungaged inflow for the period October 1982 to October 1983. Results of both Spring Creek and well 5S-31E-27ABA1 regression equations were satisfactory in that the difference between regression equation estimates and the ungaged inflow as compiled by water budgets was generally less than 2 percent of the combined monthly discharges of the Snake River near Blackfoot and at Neeley (fig. 10). Water levels in well 4S-33E-3CBB2 were less satis­ factory than the other regression variables because water levels in 1983-84 tended to be higher than 1978-82 levels. Lag times between water-level changes in wells and discharge peaks in Spring Creek were determined by comparing well and discharge hydrographs. Maximum water levels in DIFFERENCES IN DISCHARGE, IN CUBIC FEET PER SECOND

o V) Wi O o 0 0 0 o O O O O

o

(TO* s (O CO ro O n o3 oH-

53 3a ^ no « ^ P P 2 no s s ^ ffi r PO m i cro «->- tn o X o

*« 2

CO p 2 § U)

£ ;fl £;

HO ^d g

' II

O well 5S-31E-27ABA1 seemed to lag peaks in Spring Creek by an average of about 9 days. Water levels in well 4S-33E-3CBB2 seemed to peak at nearly the same time as Spring Creek discharge. The regression relations for Spring Creek were devel­ oped using monthly data but also can be used to calculate daily inflow, although some additional error is introduced. As previously determined, ground-water discharge is about 94 percent of the average ungaged inflow to American Falls Reservoir, and daily changes in ground-water discharge are probably small. Most of the change in daily ungaged inflow (and the source of additional error) is due to runoff from precipitation and irrigation-return flow that is much more variable than the monthly averages used to develop the regression equations. Figure 11 shows that increases in daily discharge of Spring Creek above base flow generally correspond to precipitation recorded at nearby weather stations. Increases in discharge do not always correspond to precipitation at recording stations because rainstorms are scattered geographically. Use of Spring Creek daily discharge as the independent variable in the regression equation would account for some of the regression error introduced by computing daily ungaged inflows with an equation developed from monthly data. Further reduction in daily variances could be made by measuring flow in the larger ungaged tributaries (Bannock Creek and Ross Fork) and drains (Crystal and Aberdeen Wasteways).

Reservoir Storage Errors in determining reservoir storage change can be introduced by (1) insufficient or poor stage data for the reservoir, and (2) an inaccurate reservoir-capacity table. The first is likely to have the greater effect in computa­ tion of daily water budgets because 0.01 ft in stage will cause a change larger than the RMSE. To more accurately define the stage of American Falls Reservoir, three addi­ tional gages were installed in November 1982: at Sterling in the northernmost part of the reservoir, along the western shore near Aberdeen, and along the eastern shore in Seagull Bay (fig. 3). The area-capacity table (U.S. Bureau of Reclamation, 1979) was used to determine storage. To determine change in storage, average daily change in stage of all gages was multipled by the average surface area obtained from the area-capacity table. Two daily water budgets were compared for determination of ungaged inflow. For one water budget, the change in reservoir storage was determined using only data from the gage at American Falls. For the other, change in storage

29 l.V i i (f) Fort Hall UJ 5 0.5 Z Z 0 '..I, I, M 1 ll ll 1 1 1 , , , II 1 , , 1 I z"

0 1.5 ' ' 1- Pocatello 1- 1.0 - - a. uj 0.5 - - a. 0 ,1 1,1 I ll i ll i i i I 1 1 1 1 I 1 oW

m UJ O Spring Creek increase in daily discharge! > 7 ft 3/B 3 UJ Z a: "-O 20 - - < y u III 10 ~ I - o 1 March 1 April ' May ' June July ' August ' September ' 1983

Figure 11. Increases in daily discharge of Spring Creek above base flow and recorded precipitation at Fort Hall and Pocatello, March 1 to September 30, 1983. was determined using the best combination of data from the three additional gages. A comparison of standard errors of gains computed using various combinations of the four reservoir stage gages is shown in table 5. The greatest decrease in standard error was achieved using data from the gages at Sterling, Aberdeen, and American Falls. Gains or losses in the water budget determined using data from these gages were compared with estimated ungaged inflow computed from the Spring Creek regression equation for the period March to September 1983 (fig. 12). Standard errors of the daily estimates for each month from March through September are shown in figure 13. Greatest decreases in standard error resulting from use of the three gages were in September when stage in the reser­ voir was rapidly decreasing. Although stage also was decreasing in August, little decrease in standard error was evident. High winds in mid-August may have caused fluctua­ tions in the recorded stage, particularly at the Sterling gage.

Reservoir Leakage Leakage from American Falls Reservoir contributes some spring discharge to the Snake River between Neeley and Minidoka (Mundorff, 1967, p. 38). Spring discharge increased after storage began in the reservoir. Although water levels in wells at the southwestern end of American Falls Reservoir (fig. 14) increase from February to May when stage in the reservoir increases (fig. 15), the lack of correlation between ground-water levels and reservoir stage at other times may indicate that leakage is small. Water levels in well 7S-30E-14DCC1 along the west side of American Falls Reservoir indicate a locally perched aquifer that is recharged by percolation of irrigation water. Water levels in the well increase from May to September and decline thereafter. Water levels in wells 7S-30E-15AAA1 and 7S-30E-28BBC1, west of the Aberdeen- Springfield Canal, increase from February to May and de­ crease from May to July. Water-level trends in wells 7S-30E-13DCA1, 7S-30E-24DDC1, and 7S-30E-26DDD1 near the southwestern end of the reservoir were like those in wells 7S-30E-15AAA1 and 7S-30E-28BBC1. Water levels in wells along Lake Channel and stage in Bonanza Lake also follow the same pattern (fig. 16). The rise seems to diminish down- gradient from the reservoir. Declines in water levels probably are accelerated by ground-water pumping for irrigation. On the basis of water levels in wells near the reservoir, the effect of recharge from reservoir leakage on water levels is masked by the effect of ground-water pumpage. Increases and decreases in leakage owing to alternating high and low reservoir stages are not apparent.

31 Table 5. Comparison of standard errors of gains computed from combinations of reservoir stage gages (April 15 to September 15, 1983) [x f staqe gages used in the computation]

in rH rH £*i

X 1,231 X X X X 809 34 X X 908 26 X X X 781 37 X X X 924 25 X X X 848 31

32 7,000

Estimated ungaged G ain or loss from inflow from Spring Creek daily water budgets regression equation

Average ungaged inflow from Spring Creek regression equation

CO CO

Average gain or loss from daily water budgets

-2,000 March September

Figure 12.~Comparison between water-budget gain or loss and estimated ungaged inflow to American Falls Reservoir based on Spring Creek discharge, March 1 to September 30, 1983. 2,000

Reservoir storage determined from gage at dam

1,800 Reservoir storage determined from three gages

1,600 O u UJ

£ 1,400 0.

UJ UJ u_ y 1,200 CQ u tx 1,000 O

UJ Q o: 800

600

400

200 March April May June July August September I1983 : I

Figure 13. Standard errors of two daily water-budget estimates of ungaged inflow to American Falls Reservoir from estimates derived from regression equations that relate ungaged inflow to Spring Creek discharge.

34 EXPLANATION

-A" STAGE-MEASUREMENT STATION 113°00 ^ DISCONTINUED GAGING STATION

A GAGING STATION 13DCA1 OBSERVATION WELL AND NUMBER

43°45'~-

w Ol

10 KILOMETERS

Figure 14.--Lake Walcott reach of the Snake River. 4,370

UJ UJ _J .7S-30E-14DCC1 < 4,360 UJ

UJ > Stage of American Falls Reservoir O CD

I- 4,350 UJ UJ

Z g I- 4,340 0> _jUJ ill

4,330 Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. 1982 1983

Figure 15. Stage of American Falls Reservoir and elevation of water levels in downgradient wells. 4,260

4,255

LLI

< 4,250 LU C/5 LU Stage of Bonanza Lake > 4,245 O ^Missing records CQ

H 4,240 LU LU X LU 9S-29E-4BCA1- 2 4,215 w 4 O H 4,210

LU 4,205

4,200 Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. 1982 1983

Figure 16. Stage of Bonanza Lake and water levels in nearby wells. Lake Walcott Minidoka Dam, completed in 1909 to divert water for irrigation projects both north and south of the Snake River, is about 40 mi downstream from American Falls Dam. Back­ water from Lake Walcott, behind Minidoka Dam, may extend about 33 mi upstream. Lake Walcott has a usable storage capacity of 107,000 acre-ft between elevations 4,236 and 4,246 ft. A powerplant at the dam generates about 90,000 MWh annually (Heitz and others, 1980, p. 130). Discharge through the powerplant and spillways is inflow to Milner Lake. Water budgets were prepared for the reach of the Snake River from Neeley to Minidoka. Inflow to the reach is discharge of the Snake River at Neeley; outflow from the reach is discharge of the Snake River near Minidoka plus diversions to the North Side and South Side Minidoka Canals. Gains or losses in the reach are the difference between inflow and outflow adjusted for precipitation, storage change, and evaporation in Lake Walcott. Average annual gain from 1951 to 1983, determined using the water budgets, was 245 ft 3 /s. Figure 17 shows the 5-year moving average (four previous years plus current year) gain in the reach from 1953 to 1983, compared with cumulative departures from mean discharge of Goose Creek near Oakley (fig. 1). From this comparison, it appears that the high gain from 1970 to 1983 is related to high runoff from tributaries. Goose Creek discharge is assumed to represent surface-water runoff from tributaries to the Snake River between Neeley and Minidoka. Inflow between Neeley and Minidoka gages from the north side of the river is entirely from springs and ground-water seepage. Gifford Springs (fig. 13) is the largest group of these springs. Stearns and others (1938, p. 153) estimated Gifford Springs discharged 25 to 35 ft3 /s prior to irriga­ tion of the Aberdeen-Springfield tract in about 1900. Backwater from normal water levels in Lake Walcott has prevented further discharge measurements of these and several other springs. Discharges of Rueger and Mary Franklin Mine Springs increased about 70 and 55 percent, respectively, after storage in American Falls Reservoir began in 1926. The average of eight discharge measurements of Rueger Springs in 1925-26 was 11.8 ft 3 /s, and the average of three discharge measurements in 1928 was 19.9 ft 3 /s (Stearns and others, 1938, p. 151). The average of 25 discharge determinations of Mary Franklin Mine Springs in 1925-26 was 5.8 ft 3 /s, and the average of 12 determinations in 1928 was 9.0 f t 3 /s (Stearns and others, 1938, p. 153). Mundorff (1967, p. 40) estimated that spring discharge and

38 500

* Cumulative departure (Goose Creek)

WATER YEAR

Figure 17. Comparison between 5-year moving average gain to Lake Walcott and cumulative departures from mean discharges of Goose Creek near Oakley, 1953-83.

39 seepage to the Snake River were 80 to 100 ft 3 /s before construction of American Falls Dam and that the average discharge increased 15 to 20 ft 3/s after construction. For the Neeley gaging station, 30 ft 3 /s are added to each discharge measurement to make them compatible with histor­ ical measurements made at an older site below the springs. Therefore, actual net spring discharge to the reach is about 65 to 90 ft 3 /s (95 to 120 ft 3 /s minus 30 ft 3 /s). Inflow between Neeley and Minidoka gages from the south side of the river is largely from Rock Creek basin (fig. 13). Williams and Young (1982, p. 20) estimated the average surface-water flow from Rock Creek to the Snake River was 16,500 acre-ft/yr, or 23 ft 3 /s. They also estimated that 51,000 acre-ft/yr, or 70 ft 3 /s, moves as underflow from the basin to the Snake River (p. 38) . Rock Creek near American Falls discharged 85 percent of its annual flow from November to April in water years 1979 and 1980. Mean discharges for these 2 years are 27.4 and 39.8 ft 3 /s. Maximum daily mean discharges are 1,760 ft 3 /s for the November through March period and 173 ft 3 /s for the April through October period. Mean discharges are 126 ft 3/s for November through March and 13 ft 3 /s for April through October. Streamflow in other tributaries, including tributaries to the lower part of the Raft River, generally exhibits seasonal trends similar to trends for Rock Creek, and little or no flow enters the Snake River in the summer. Yield from Warm Creek, Falls Creek, and other tribu­ taries (a total drainage area of about 60 mi 2 ) is esti­ mated to be equal to the rate of yield from Rock Creek basin, or 0.29 (ft 3 /s)/mi. Discharge from the upper Raft River basin is high April through June, but part of this discharge does not reach the Snake River because of diver­ sions and seepage. The average annual surface-water flow reaching the Snake River is about 10 ft 3 /s. Ground-water outflow from the Raft River basin averages 80,000 acre- ft/yr, or 110 ft 3/s (Walker and others, 1970, p. 112). Nace (1961, p. 103) indicated that at least some of the ground water does not discharge to the Snake River but passes beneath it. Water-table contour maps for the area indicate that most of the ground-water movement is westward. As a result of pumpage for irrigation, water-level declines in the Raft River basin from 1971 to 1982 (Young and Norvitch, 1984, p. 11) have reduced the hydraulic gradient, and ground-water underflow to the Snake River from the Raft River basin is probably much less at present than the 80,000 acre-ft estimated by Walker and others (1970). Table 6 lists sources of inflow to Lake Walcott and their likely contributions. The north-side springs and seeps are assumed to be at the high end of the estimated range in discharge. The sum of north-side spring discharge and south-side surface- and ground-water inflow is equal to

40 Table 6. Inflow to Lake Walcott, 1951-83

Average annual discharge Sources of inflow (cubic feet per second)

North-side springs and seepage 90 Rock Creek basin: Surface water 23 Ground water 70 Water yield from small tributary basins 42 Raft River basin: Surface water 10 Ground water (residual) 10 Total (set equal to average gain) 245

41 the gain determined by water-budget analysis (245 ft 3 /s) when ground-water inflow from the Raft River basin is 10 ft3 /s. Because part of the ground-water outflow from the Raft River basin is to the west and to the north under the Snake River, additional inflow is likely to be small and, consequently, leakage from the reservoir would be small. Stearns and others (1938, p. 197) reported that seepage losses from Lake Walcott in the first 52 months of its existence were about 1.4 million acre-ft, or an average flow of 450 ft 3/s. Stearns also reported that, in the following 52 months, the losses averaged 140 ft3 /s. Present losses appear to be much less than 140 ft 3 /s. The decrease in losses can be attributed to sediment sealing the bottom of Lake Walcott. Discharge-measurement errors of a few percent can influence whether water-budget calculations show a gain or loss in the Neeley to Minidoka reach. Five percent of the average flow of the Snake River at Neeley (7,256 ft 3 /s) is 360 ft 3 /s compared to the average gain of 245 ft 3 /s. To obtain a more accurate estimate of gains to the reach, monthly mean discharge data from 33 months during the period 1935-82 when combined inflow and outflow averaged less than 2,000 ft 3 /s were used in water budgets to determine gain. The 33 months selected were all in the nonirrigation season from November to March. The combined inflow and outflow of less than 2,000 ft 3/s was selected because residuals would be a larger percentage of the total. Mean gain to Lake Walcott for the 33 months was 203 ft 3 /s; the standard deviation was 92 ft 3 /s (fig. 18). Although only the period November to March was included in the analysis, the mean gain also should be representative of that from April to October because high tributary flows in April and May would be balanced by low flows in July, August, and September. For the data examined, the smallest gain was 51 ft 3/s in November 1961, and the largest gain was 511 ft3 /s in January 1980. This large gain was verified by the monthly mean discharge for Rock Creek near American Falls, which was 153 ft 3/s. Leakage from Lake Walcott would average 40 to 50 ft3 /s if 203 ft3 /s is assumed to be the average gain. This amount of leakage seems reasonable, given the leakages that oc­ curred in the first several years of storage. Improvements in the water-budget analysis could be made by additional data collection. Ground-water discharge could be quantified by measuring discharge in the Snake River when reservoir stage and river discharge were low. Miscellaneous discharge measurements from other sources of surface-water runoff could be correlated with the discharge record of Rock Creek.

42 MAXIMUM MONTHLY MEAN GAIN MINIMUM MONTHLY MEAN GAIN

MONTHLY MEAN GAIN

STANDARD DEVIATION OF MONTHLY MEAN GAINS FOUR PERCENT OF AVERAGE SNAKE RIVER DISCHARGE

0 200 400 600 DISCHARGE, IN CUBIC FEET PER SECOND

Figure 18.~Statistical summary of gains to Lake Walcott during low flow in the Snake River, 1935-82. Milner Lake Milner Dam, completed in 1905, was one of the first dams on the Snake River. Most diversion works and canals were completed by 1910. Water is diverted from headgates at the dam and is pumped from Milner Lake to irrigate downstream land on both sides of the Snake River. Above the dam, most of Milner Lake is confined in a narrow, 34.5-mi long canyon. Because of the long canyon, reservoir elevation varies throughout the reach. Consequently, a step-backwater analysis was used to develop a reservoir- capacity table. Water budgets were calculated for the reach of the Snake River from Minidoka to Milner (fig. 19) by assuming ungaged inflow to be the difference between inflow and outflow and adjusting for the change in reservoir storage. Inflow is discharge of the Snake River near Minidoka; outflow is discharge of the Snake River near Milner plus several diversions. During parts of August and September 1983, most of the Snake River discharge was diverted (fig. 20). In some years, leakage from the dam constitutes the only flow past the gaging station near Milner during parts of August and September. Neither precipitation on nor evaporation from Milner Lake are included in the water budget because of the narrowness of the reach and its relatively small surface area. Instead, gains from precipitation and losses from evaporation are included as part of the net gain or loss in the reach.

Gain and Loss Analysis Gains and losses in the reach are small relative to the amount of water flowing through the Milner Lake reach, so that discharge-measurement errors of a few percent are likely to be larger than the actual gain or loss much of the time. However, discharge-measurement errors may balance out when averages are computed for longer time periods. Mean monthly gains and losses in Milner Lake for water years 1975-83 were determined from water budgets (contents in Milner Lake have been determined only since 1975). Results indicate that highest gains were in September and October (more than 400 ft 3 /s) and that losses of about 100 ft 3 /5 may have occurred in May (fig. 21). Standard deviations are generally near +4 percent of the monthly mean discharge of the Snake River near Minidoka (fig. 22) and probably represent discharge-measurement errors rather than actual variations in gain or loss.

44 113° 00'

43°45'--

EXPLANATION

STAGE-MEASUREMENT STATION

GAGING STATION

OBSERVATION WELL 23DBA1 AND NUMBER 3

03

43°30

10 KILOMETERS

Figure 19. Milner Lake reach of the Snake River. 14,000

12,000

10,000 u HI

a.HI 8,000 HI U_ y CQ D U 4" 6,000 0) u (D

Z 4,000

2,000

April May June July August September October 1983

Figure 20. Outflow from Milner Lake. ore*3 c GAIN AND LOSS, IN CUBIC FEET PER SECOND

3S O

VO 3 -J to Y1 * u>00 a.3 o" to to 5* 600

Standard deviation of gain or loss

Four percent of mean monthly discharge, 500 Snake River near Minidoka uO SLU 40° HI 0.

HI y 300 CQ

LU (D (£ 200 m\

U (/)

100

Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept.

WATER YEAR

Figure 22. Standard deviations of gains or losses in Milner Lake and possible discharge-measurement errors in mean monthly discharge, 1975-83. Total surface-water inflow to the Snake River between Minidoka and Milner was measured seven times between March 1980 and March 1981 (table 7). Most gains are from irrigation-return flow, but snow- melt and rainstorm runoff can produce flows of several hundred cubic feet per second for short periods. Measured surface-water inflow during the 1980 irrigation season was greater than the average gain from 1975 to 1983. Average gain or loss computed from a few years of data could be highly inaccurate as an estimate of short-term values. Although surface-water inflow provides most of the gain from Minidoka to Milner, water levels in well 9S-25E-23DBA1, south of the Snake River at the upper end of Milner Lake, indicate that the hydraulic gradient is toward Milner Lake. Seepage from ground water then would contribute to gains in the reach, at least when ground-water levels are highest near the end of the irrigation season (fig. 23). Most of the time, however, ground-water levels on both the north and south sides of the Snake River are at a lower elevation than the lake level, and there may be some net losses to ground water. The average annual gain in Milner Lake from 1951 to 1983 was 290 ft 3 /s. Figure 24 shows the 5-year moving average of annual mean gain in Milner Lake and water levels in well 9S-25E-23DBA1. Since 1976, gain in the reach has declined. This may be due to about a 10-percent decrease in diversions from the Snake River since 1975 and the use of sprinkler irrigation systems, which are generally more efficient than flood irrigation. Water levels in well 9S-25E-23DBA1 also have declined, which would be anticipated with reduced gains in Milner Lake. Monthly mean water-budget gains or losses were deter­ mined for 14 months from 1977 to 1982 (fig. 25) when discharge in the Snake River was low. Four percent of the monthly mean discharge in the Snake River near Minidoka is shown as a reliability indicator; however, water-budget errors could be greater. Standard deviation of daily gain or loss is another reliability indicator. It is apparent that even for low flows, daily gains and losses determined by water budgets are highly inaccurate. From the seemingly more reliable months, average gain during the nonirrigation season probably would range from 100 to 200 ft 3/s. The large gain for February 1982 may be attributed to peak discharges at gaging stations on Raft River, Goose Creek, and Rock Creek on February 16.

49 Table 7. Surface-water inflow to the Snake River between Minidoka and Milner

Surface-water inflow (cubic feet Date per second)

3-26-80 20 5-22-80 235 7-22-80 268 9-16-80 563 11- 5-80 75 1-22-81 6 3-10-81 14

SO 4,135

4,133

UJ LU U. > 4,131

O c/) I- "J 4,129 >o HI OQ -I < UJ 4,127

4,125 I I I I I I I I ONDJFMAMJJAS WATER YEAR

Figure 23.--Water levels in well 9S-25E-23DBA1, water year 1983.

51 OQ

GAIN TO REACH, IN CUBIC FEET PER SECOND

o O O o o o O O o o 1953

5 3 «^ + o3 1963

H m en to m 2: 1973 vo

w 1983 n DEPTH TO WATER, IN FEET OQ 5*

N-« 9 EXPLANATION

MONTHLY MEAN GAIN OR LOSS IN MILNER LAKE

FOUR PERCENT OF DISCHARGE, SNAKE RIVER NEAR MINIDOKA

STANDARD DEVIATION OF DAILY GAIN OR LOSS IN MILNER LAKE 800

700

Q § 600 U HI

K HI 500 D.

HI 400 U. y CO Ul D 30° CO U Z HI* 200

I y 100

-100 I Oct. Nov. Dec. Jan. Feb. Nov. Dec. Jan. Feb. Mar. Nov. Dec. Jan. Feb.

Figure 25. Monthly mean gain or loss in Milner Lake for selected months, 1977-82. Reservoir Storage In 1974, a storage-capacity table was developed for Milner Lake that considered both stage at the dam and discharge at the upper end of the reach near Minidoka. However, discrepancies in water-budget calculations during this study made it apparent that a better definition of the backwater curve and cross-sectional area of canyon was required. Data obtained during 1983 and 1984 better defined the capacity table for Milner Lake. Additional reservoir stage measurements were obtained at Burley and near Rupert (fig. 19). Data for 14 cross sections in Milner Lake and 1 cross section at the measurement site near Minidoka (fig. 19), reservoir stage at the dam and near Rupert and Burley, and discharge of the Snake River near Minidoka were used to redefine backwater curves for a range of reservoir stages and discharges. The U.S. Geological Survey step-backwater computer program, E 341 (Shearman, 1976), was used to determine water-surface elevation at each of the 15 cross sections. Water-surface elevations for the Burley and Rupert gaging sites were compared with interpolated eleva­ tions from the computed backwater curve elevations at cross sections 11-12 and 13-14, respectively (fig. 19). To obtain a proper fit of the backwater curve, rough­ ness coefficients ("n" values) were adjusted until water- surface elevations were within a few hundredths of a foot of the recorded stages at Burley, near Rupert, and near Minidoka. Roughness coefficients normally vary inversely with discharge and, in the upper subreaches, n values ranged from 0.029 to 0.023 in the main channel at discharges from 6,000 to 16,000 ft 3 /s. In the lower subreaches, n values ranged from 0.016 to 0.020 at these same discharges. The low n values are attributed to water flowing through the reservoir, where boundary friction is reduced. Water-surface profiles in Milner Lake for a constant reservoir elevation of 4,134 ft at Milner and two discharges are shown in figure 26. Profiles for a constant inflow of 9,000 ft 3 /s and two reservoir elevations at Milner are shown in figure 27. The new storage-capacity table for Milner Lake devel­ oped from corresponding surface areas and profiles is given in table 8. Backwater curves were determined and capacities were computed for each 1,000 ft 3 /s of discharge of the Snake River near Minidoka (below 1,000 ft 3 /s, 100 ft 3 /s increments were used) and each 1-ft change in stage of Milner Lake at the dam. For all other combinations of discharge and stage, storage can be interpolated between values given in the table.

54 4,150

4,140 Elevation at EXPLANATION Milner, 4,134 ft UJ > UJ WATER-SURFACE PROFILE WHEN 4,130 DISCHARGE AT MINIDOKA IS 24,000 ft 3/s. TOTAL STORAGE, 46,526 ACRE-FEET (f)UJ 4,120 O CD WATER-SURFACE PROFILE WHEN DISCHARGE AT MINIDOKA IS H 4,110 900 ft 3/s. TOTAL STORAGE, UJ UJ 36,273 ACRE-FEET

4,100 Z O h- THALWEG OF SNAKE RIVER en 4,090 ui g UJ _i UJ

4,080

4,070 0 20 40 60 80 100 120 140 160 180 200 DISTANCE ABOVE MILNER DAM, IN THOUSANDS OF FEET

Figure 26. Water-surface profiles in Milner Lake for a constant reservoir elevation at Milner and two discharges. 4,140

4,130 EXPLANATION _j LU > LU Discharge near Minidoka, 9,000 ft 3/s WATER-SURFACE PROFILE WHEN < 4,120 RESERVOIR ELEVATION IS LU 4,134 FEET. TOTAL STORAGE, C/5 39,116 ACRE-FEET 111 > § 4,110

I- WATER-SURFACE PROFILE WHEN Lll 111 RESERVOIR ELEVATION IS 4,100 4,125 FEET. TOTAL STORAGE, 21,897 ACRE-FEET

4,090 0> THALWEG OF SNAKE RIVER Ill _J 111 4,080

4,070 20 40 60 80 100 120 140 160 180 200

DISTANCE ABOVE MILNER DAM, IN THOUSANDS OF FEET

Figure 27. Water-surface profiles in Milner Lake for a constant inflow and two reservoir elevations at Milner. Table 8« Storage capacity table for Milner Lake [Storage capacity in acre-feet]

Discharge of Snake River near Minidoka Gage height of Milner Lake at the dam (datum 4,122,.5 feet above sea level) (cubic feet per second) 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5 10.5 11.5

500 11,222 12,921 14,380 15,995 17,873 20,102 22,706 25,636 28,861 32,392 36,130 600 11,444 13,140 14,577 16,195 18,063 20,276 22,837 25,739 38,949 32,443 36,161 700 11,660 13,358 14,789 16,383 18,266 20,436 22,967 25,841 29,205 32,494 36,169 800 11,882 13,571 14,981 16,563 18,444 20,590 23,110 25,942 29,103 32,5,67 36,234 900 12,091 13,781 15,178 16,758 18,603 10,757 23,238 26,041 29,181 32,626 36,273 1,000 12,297 13,989 15,359 16,935 18,763 20,903 23,363 26,165 29,260 32,694 36,312 2,000 14,081 25,608 16,831 18,282 20 f 046 22,094 24,441 27,061 30,036 33,168 36,778 3,000 15,635 16,370 17,523 18,961 20,640 22,631 24,928 27,511 30,379 33,589 37,004 4,000 17,032 17,264 18,373 19,735 21,349 23,259 25,496 28,015 30,852 33,930 37,316 5,000 17,644 18,320 19,320 20,679 22,182 24,048 26,201 28,621 31,394 34,431 37,671 6,000 17,833 18,463 19,489 20,845 22,338 24,196 26,344 28,758 31,517 34,543 37,764 en 7,000 18,490 19,350 20,356 21,652 23,104 24,873 26,956 29,306 32,107 34,975 38,154 s] 8,000 20,037 20,841 21,957 23,076 24,440 26,091 28,004 30,289 32,885 35,673 38,739 9,000 20,975 21,897 22,794 23,883 25,205 26,822 28,726 30,924 323,447 36,110 39,116 10,000 22,721 23,456 24,335 20,353 26,604 28,140 29,963 32,069 34,348 36,985 39,870 11,000 23,669 24,400 25,249 26,232 27,461 28,902 30,640 32,651 34,983 37,510 40,341 12,000 24,590 25,752 26,576 27,530 28,673 29,281 30,997 33,007 35,299 37,798 40,827 13,000 24,853 26,098 26,910 27,858 28,992 29,970 31,666 33,630 35,872 38,435 41,001 14,000 25,667 26,434 27,235 28,185 29,303 30,688 32,326 34,250 36,413 38,909 41,491 15,000 26,449 27,700 28,483 29,391 30,479 31,098 32,967 34,845 37,001 39,418 42,275 16,000 27,204 27,932 28,722 29,634 30,697 32,023 33,613 36,011 37,564 39,935 42,754 17,000 27,552 29,242 29,988 30,864 31,916 32,696 34,238 36,036 38,106 40,431 43,226 18,000 28,680 29,402 30,160 31,042 32,080 33,374 34,854 36,626 38,669 40,933 43,714 19,000 29,409 30,040 30,774 31,633 32,649 33m983 35,463 37,190 39,185 41,426 44,177 20,000 30,103 30,748 31,467 32,318 33,309 34,604 36,061 37,762 39,735 41,921 44,651 21,000 31,843 32,035 32,723 33,546 34,522 35,710 37,107 38,738 40,632 42,760 45,129 22,000 32,093 32,712 33,405 34,211 35,157 36,333 37,730 39,320 41,164 43,273 45,589 23,000 32,777 33,380 34,062 34,857 35,794 36,943 38,311 39,885 41,741 43,779 46,063 24,000 33,441 34,717 35,372 36,110 37,046 37,567 38,880 40,465 42,264 44,308 46,526 25,000 34,790 35,370 36,010 36,775 37,667 38,767 40,072 42,557 43,333 45,252 46,732 26,000 35,448 36,014 36,657 37,407 38,307 39,386 40,658 42,149 43,856 46,077 48,196 27,000 36,120 36,645 37,285 38,043 38,928 39,997 41,245 42,715 44,390 46,579 48,684 28,000 36,724 37,268 37,915 38,640 39,541 40,594 41,799 43,248 45,285 47,112 49,177 29,000 37,340 37,881 38,509 39,259 40,142 41,167 42,423 44,152 45,797 47,627 49,675 30,000 37,963 38,897 39,491 40,252 41,100 42,137 43,351 44,726 46,300 48,197 50,155 SUMMARY Agriculture is the most important industry on the Snake River Plain and, because of a semiarid climate, its success relies on irrigation. Determinations of gains and losses for American Falls Reservoir, Lake Walcott, and Milner Lake on the Snake River are necessary to allocate water for irrigation on the basis of established water rights. Gains and losses estimated from daily water budgets are variable, owing to inadequate determination of (1) changes in reser­ voir storage, (2) streamflow, (3) lake-surface precipi­ tation, and (4) lake-surface evaporation. The combined discharge of the Snake River near Black- foot and the Portneuf River at Pocatello accounts for about 65 percent of the inflow to American Falls Reservoir. The remainder is spring discharge, ground-water seepage, small- tributary streamflow, irrigation-return flow, and precipita­ tion on the reservoir. Precipitation, which is estimated from nearby recording stations, contributes about 0.5 percent of the average inflow. Outflow is measured in the Snake River at Neeley and at two diversions. Evaporation from the reservoir is estimated from pan data. Neither precipitation nor evaporation can be estimated with much reliability, but both can have a significant effect on a daily water budget. Average annual ungaged inflow to American Falls Reservoir from spring discharge, ground-water seepage, small tributaries, and irrigation-return flow from 1912 to 1983 was 2,690 ft 3 /s. About 94 percent of the ungaged inflow, or 2,540 ft 3 /s, is the average ground-water discharge (springs and seepage). Seasonal fluctuations of ungaged inflow seem to have increased after storage in American Falls Reservoir began in 1926. Bank storage in the reservoir and increased seasonal fluctuations in ground- water levels as a result of irrigation are probable causes. Ungaged inflow to American Falls Reservoir generally can be estimated more accurately by correlation with ground- water levels in wells or measured spring discharge than by water budgets based on discharge measurements alone. However, analysis indicates that the discharge of Spring Creek, the largest ungaged tributary to the reservoir, is a better indicator of ungaged inflow than ground-water levels. Most of the flow in Spring Creek, like most of the ungaged inflow, comes from springs. However, flow in Spring Creek also includes occasional runoff from precipitation that is not readily represented by water levels in wells. Leakage from American Falls Reservoir is small, as indicated by water levels in wells downstream from the reservoir and water budgets for the Snake River between

58 Neeley and Minidoka. Although water levels in wells at the southwestern end of American Falls Reservoir and along Lake Channel increase from February to May when stage in the reservoir increases, it cannot be shown that the changes are due to reservoir leakage. Declines in water levels in wells from May to September probably are accelerated by ground- water pumping for irrigation. The continued high reservoir stage during this period seems to have no effect on these declines. Spring discharge to the Snake River between Neeley and Minidoka increased by about 25 percent, or about 15 to 20 ft 3/s, after storage began in American Falls Reservoir in 1926. Water budgets generally were balanced when 65 to 90 ft 3/ s of spring discharge was assumed. Verification of spring discharge could be made by measuring the Snake River below the springs when stage in Lake Walcott is low and river flow is low. Annual gains to Lake Walcott from 1951 to 1983, deter­ mined using annual water budgets, averaged 245 ft 3 /s. The tendency for higher gains in the latter part of the 1951-83 period appears to be associated with high runoff from tributaries. An average of 203 ft 3 /s was determined by a water-budget analysis of 33 months when Snake River discharge was low and normal discharge-measurement errors would not have significant effect. The total of estimates of discharge from springs and seeps along the north side of the Snake River and upper Lake Walcott and streamflow and underflow from tributary drainage basins south of the River and lake is nearly equal to the 1951-83 average. However, leakage from the lower part of the reservoir would be about 40 to 50 ft3 /s using the 33-month average gain of 203 ft3 /s. Water budgets are not reliable for estimating gains and losses in Milner Lake because discharge-measurement errors are likely to be larger than the gains and losses. Annual gains in Milner Lake from 1951 to 1983, determined using water-budget analysis, averaged 290 ft 3 /s. Gains to Milner Lake are mostly from irrigation-return flow and are highest in September and October. The declines in gains since 1976 may be due to decreased diversions and increased use of sprinkler-irrigation systems. Milner Lake generally is confined to the narrow Snake River canyon, and water levels along the reservoir vary because of backwater. The backwater curve has been defined for combinations of stage at the dam and discharge of the Snake River near Minidoka. A step-backwater computer program was used to determine water-surface elevations at 15 cross sections. Roughness coefficients were adjusted until

59 surface-water elevations of the backwater curve matched recorded elevations at Burley, near Rupert, and near Minidoka. This information was used to prepare a table of storage for various combinations of inflow discharge and gage height at the dam.

60 REFERENCES CITED Barnes, H.H. Jr. , 1961, Roughness characteristics of natural channels: U.S. Geological Survey Water-Supply Paper 1849, 213 p.

Crosthwaite, E.G., 1974, A progress report on results of test-drilling and ground-water investigations of the Snake Plain aquifer, southeastern Idaho--part 3: Idaho Department of Water Resources, Water Informa­ tion Bulletin 38, 25 p. Decker, S.O., Hammond, R.E., and Kjelstrom, L.C., 1970, Miscellaneous streamflow measurements in Idaho, 1894- 1967: U.S. Geological Survey Basic-Data Release, 310 p.

Garabedian, S.P. , 1984, Application of a parameter-estimation technique to modeling the regional aquifer underlying the eastern Snake River Plain, Idaho: U.S. Geological Survey Water-Supply Paper 2278, 60 p.

Heitz, L.F., Warnick, C.C., and Gladwell, J.S., 1980, Idaho hydroelectric potential, upper Snake basins: Idaho Water Resources Research Institute, Appendix V, 177 p.

Kjelstrom, L.C., 1986, Flow characteristics of the Snake River and water budget for the Snake River Plain, Idaho and eastern Oregon: U.S. Geological Survey Hydrologic Investigations Atlas HA-680, scale 1:1,000,000, 2 sheets.

Meisler, Harold, 1958, Preliminary report on ground water in the Bonanza Lake area, Power and Blaine Counties, Idaho: U.S. Geological Survey Open-File Report, 32 p.

Meyers, J.S., 1962, Evaporation from the 17 Western States: U.S. Geological Survey Professional Paper 272-D, 100 p. Mundorff, M.J., 1967, Ground water in the vicinity of American Falls Reservoir, Idaho: U.S. Geological Survey Water-Supply Paper 1846, 58 p.

Mundorff, M.J., Crosthwaite, E.G., and Kilburn, Chabot, 1964, Ground water for irrigation in the Snake River basin in Idaho: U.S. Geological Survey Water-Supply Paper 1654, 224 p.

Nace, R.L., and others, 1961, Water resources of the Raft River basin, Idaho-Utah: U.S. Geological Survey Water-Supply Paper 1587, 138 p.

61 National Oceanic and Atmospheric Administration, 1983, Climatological data, Idaho: Asheville, N.C., monthly summaries. Newell, T.R., 1928, Segregation of water resources, American Falls basin and reservoir: Idaho Falls, Report to Water District No. 26, 127 p. 1929, Segregation of water resources, American Falls basin and reservoir: Idaho Falls, Report to Water District No. 26, 144 p. Shearman, J.O. r 1976, Computer applications for step- backwater and floodway analyses: U.S. Geological Survey Open-File Report 76-499, 103 p. Stearns, H.T., Crandall, Lynn, and Steward, W.G., 1938, Geology and ground-water resources of the Snake River Plain in southeastern Idaho: U.S. Geological Survey Water-Supply Paper 774, 268 p. U.S. Bureau of Reclamation, 1979, American Falls Reser­ voir area-capacity table: Boise, Idaho, 16 p U .S. Geological Survey, 1968-78, Water resources data for Idaho, water years 1968-78: Boise, Idaho, U.S. Geologi­ cal Survey water-data reports (published annually). Walker, E.H., Dutcher, L.C., Decker, S.O., and Dyer, K.L., 1970, The Raft River basin, Idaho-Utah, as of 1966 a reappraisal of the water resources and effects of ground-water development: U.S. Geological Survey Open-File Report, 119 p. Water District 01, 1981, Snake River and tributaries above Milner, Idaho: Idaho Falls, Annual Watermasters Report, 72 p. Williams, R.P., and Young, H.W., 1982, Water resources of Rockland basin, southeastern Idaho: U.S. Geological Survey Open-File Report 82-755, 62 p. Winter, T.C., 1981, Uncertainties in estimating the water balance of lakes: American Water Resources Associa­ tion, Water Resources Bulletin, v. 17, no. 1, p. 82-114. Young, H.W., and Norvitch, R.F., 1984, Ground-water-level trends in Idaho, 1971-82: U.S. Geological Survey Water-Resources Investigations Report 83-4245, 28 p.

62